摘要
为提升并联式混合动力汽车(parallel hybrid electric vehicle,PHEV)的燃油经济性,针对等效燃油消耗最小控制策略(equivalent fuel consumption minimum strategy,ECMS)在不同工况下适应性差的问题,以优化整车等效燃油消耗量为目标,设计基于工况识别算法的变等效因子ECMS能量管理策略。选取3类典型工况建立支持向量机分类模型,通过递归特征消除法对样本特征进行选择,采用鲸鱼算法对支持向量机进行参数优化,使用模拟退火算法分别对3类工况的ECMS等效因子进行离线全局最优求解,并分别存储于等效因子库中,通过训练好的支持向量机分类器对目标优化工况进行工况识别,不同类型的工况片段采用不同的等效因子进行转矩分配。仿真结果显示:相比于逻辑门限能量管理策略,基于工况识别算法的变等效因子ECMS能量管理策略的电池荷电状态(state of charge,SOC)变化量减少8.67%,节油率为13.11%;相比于优化前的ECMS策略电池SOC变化量减少3.47%,节油率约为6.63%。本文提出的基于工况识别算法的变等效因子ECMS能量管理策略可以有效地减少燃油消耗量,提升PHEV的整车经济性。
As for the disadvantage of poor adaptability of equivalent fuel consumption minimum strategy(ECMS)under different working conditions,the energy management strategy of variable equivalent factor ECMS based on condition recognition algorithm is designed to improve fuel economy of parallel hybrid electric vehicles(PHEV).The vehicle equivalent fuel consumption is optimized as the optimization objective.Three typical working conditions were selected to establish the SVM classification model,the sample features were selected by recursive feature elimination method,and the whale algorithm was used to optimize the parameters of the SVM.The simulated annealing algorithm was used to solve the ECMS equivalent factors of the three types of working conditions for offline global optimal solution,and were stored in the equivalent factor library respectively.The target optimized working conditions were identified by the trained support vector machine classifier.Different types of working conditions were treated with different equivalent factors for torque distribution.Compared with the logic threshold energy management strategy,the variable equivalent factor ECMS energy management strategy based on the condition recognition algorithm reduced the variation of State of charge(SOC)by 8.67%,and the fuel saving rate by 13.11%.Compared with the ECMS strategy before optimization,the SOC variation of battery was reduced by 3.47%,and the fuel saving rate was about 6.63%.The variable equivalent factor ECMS energy management strategy based on the driving condition recognition algorithm can effectively reduce the fuel consumption and improve the economy of the PHEV.
作者
张代庆
牛礼民
汪恒
张义奇
ZHANG Daiqing;NIU Limin;WANG Heng;ZHANG Yiqi(School of Mechanical Engineering,Anhui University of Technology,Ma'anshan 243032 China)
出处
《西华大学学报（自然科学版）》
CAS
2024年第3期54-63,共10页
Journal of Xihua University:Natural Science Edition
基金
先进数控和伺服驱动技术安徽省高校重点实验室开放基金资助项目(XJSK202104)
电气传动与控制安徽省重点实验室开放基金资助项目(DQKJ202204)。
关键词
并联混合动力汽车
能量管理策略
工况识别
鲸鱼优化算法
支持向量机
递归特征消除
等效燃油消耗最小
parallel hybrid electric vehicle
energy management strategy
working condition identification
whale optimization algorithm
support vector machine
recursive feature elimination
minimum equivalent fuel consumption
